# how to calculate deviation for set of values??

3 views (last 30 days)
M.Prasanna kumar on 5 Feb 2020
Commented: M.Prasanna kumar on 16 Feb 2020
No of cases = 112; (112*9=1008)
in each case i have 9 obsevations( but there are missing values in between for which matlab treats them as Nan); so total i have an array of 1008*1
i calculted mean of 9 observations for each case. Now i want to calculate deviation of each observation( 9 observations) from the mean of that particular case.
M = strcat('iva','.xlsx' ); %% read excel file , iva file consists of 1008*1 array
mean_iva = zeros(112,1); %% initialising matrix
for ii = 1:9:1008
mean_iva(ii,1)= nanmean(M1(ii:ii+8)) %% calculating mean for each case
end
B = mean_iva(mean_iva~=0); %% MEAN for all 112 cases
dev = zeros(1008,1);

Guillaume on 5 Feb 2020
You could do the same thing you've done for the mean, use a for loop.
However, a much simpler way is to reshape your input vector into 112 columns of 9 rows and simply call mean and std, no loop needed:
%M1: vector whose length is a multiple of 9
assert(mod(numel(M1), 9) == 0, 'Length of M1 must be a multiple of 9');
M1 = reshape(M1, 9, []); %reshape in rows of 9 elements
mean_iva = mean(M1, 1, 'omitnan'); %calculate the mean across the rows
std_iva = std(M1, 0, 1, 'omitnan'); %standard deviation (normalised by N-1) across the rows

M.Prasanna kumar on 7 Feb 2020
sir this is calculating standard deviation which is asixe of 1*112
but i need only deviation of each observation from the mean which will be the size of 9*112
Guillaume on 7 Feb 2020
You mean the want the difference from the mean? In this case, it's simply:
meandiff = M1 - mean_iva;
M.Prasanna kumar on 16 Feb 2020
thank you sir